Thyrotropin-Releasing Hormone

George A. Mason, James C. Garbutt, and Arthur J. Prange, Jr.

In 1969, a group led by Guillemin (7) and another by Schally (5), having worked competitively for many years, announced that the hypothalamic substance that causes the anterior pituitary gland to release thyrotropin (thyroid-stimulating hormone, TSH) is L-pyroglutamyL-L-histidyl-L-prolineamide (L-pGlu-L-His-L-ProNH2). This tripeptide is now called thyrotropin-releasing hormone (TRH) (Fig. 1).

The discovery of the chemical identity of TRH verified the venerable theory that the brain can influence the anterior pituitary gland by means of hormones. It validated the discipline of neuroendocrinology: If neurons secrete hormones, they can properly be regarded as transducers between two of the great communication systems of the organism. The discovery revived interest in the role of peptides in the nervous system. Guillemin and Schally each received one-quarter of the 1977 Nobel Prize for Physiology or Medicine. Half the prize was awarded to Yalow for her contribution to the development of the radioimmune assay as a system for the detection of minute amounts of biological substances, including peptides.

TRH was not the first peptide found in the central nervous system (CNS). In 1931, von Euler and Gaddum (86) discovered substance P, though its identity as an undecapeptide was not revealed until later. In any case, there was little interest in peptidology until an endocrine target for the brain (the anterior pituitary) was posited and until the chemical nature of brain hormones could reasonably be considered to be peptidergic. Furthermore, neither Guillemin nor Schally was the first to propose that the hypothalamus might regulate the anterior pituitary by secreting substances into the portal venous system that connects them. Others had made such suggestion; among them, the physiologist Geoffrey Harris is usually credited with formulating the concept. Attempting to prove its accuracy became his life's work (24).

The discovery of TRH led quickly to the discovery of other small peptides that regulate the secretion by the anterior pituitary of its tropic hormones. Such regulation is usually accomplished by increasing or decreasing stimulation. However, growth hormone appears mainly to be regulated by increasing or decreasing inhibition, though a growth-hormone-releasing factor has been identified (55). The peptide substances of which TRH was the first to be identified are often grouped as hypothalamic hypophysiotropic hormones. This term, while accurate, is insufficient. Many of the substances are found outside the hypothalamus. If activity follows anatomy, then they may be expected to exert nonhypophysiotropic functions. Indeed, this is the case (63).

The distribution of TRH has been studied extensively in the mammalian central nervous system and some peripheral tissues (see ref. 27 for review). Provided here is a brief description of the distribution of TRH in the rat; however, in some structures considerable species-specific variation exists.

TRH-immunoreactive cell bodies are distributed throughout the CNS of the rat (27). They are prominently clustered in the glomerular layer of the olfactory bulbs, piriform and entorhinal cortices, hippocampus, amygdala, nucleus accumbens, and olfactory tubercle. TRH-positive cells are scattered in the corpus striatum, the bed nucleus of the stria terminalis, the septohypothalamic nucleus, and the bed nucleus of the diagonal band of Broca. More caudally, TRH neurons are concentrated in the parvocellular portion of the paraventricular nucleus and in the periventricular area. Such neurons are also found above the optic chiasm, in the preoptic area, the lateral aspects of the anterior hypothalamic nucleus, the perifornical area, the supraoptic nucleus, and, less prominently, in other hypothalamic nuclei. Numerous TRH-positive cell bodies occur in the periaqueductal central gray; fewer are found in the substantia nigra, the ventrolateral lemniscal nucleus, and the roots of the trigeminal nerve. In the medulla, TRH cells occur in various raphe nuclei, the external cuneate nucleus, the dorsal vagal complex, and the area postrema. They are also prominent in lamina II–III of the dorsal horn of the spinal cord. In addition, major fiber tracts are located in the lateral septum, hypothalamus, the subiculum–amygdalohippocampal area, and the ventral horn of the spinal cord.

Outside the CNS, TRH immunoreactivity is observed in pancreatic beta cells, but levels decline sharply after birth, except in hypothyroidism (27). It appears that TRH may have an important prenatal function in pancreas. TRH is also present in various parts of the gastrointestinal tract, where it affects motility, acid secretion, and absorption of sugars.

Of recent interest are immunohistochemical and immunocytochemical studies demonstrating coexistence of TRH in neurons with one or more other neuroactive substances including serotonin, substance P, the enkephalins, dopamine, neuropeptide Y, histamine, and growth hormone (see Table 1).

The coexistence in certain neurons of gamma-aminobutyric acid (GABA) (35), cholecystokinin (54), and proctolin (27) with serotonin, which has been shown to coexist with TRH (27, 68) in a large portion of medullary bulbospinal neurons, suggests that these neuropeptides may also coexist with TRH in some neuronal populations. The coexistence of various neurotransmitters and neuromodulators is of great theoretical interest. It is clear that the presence in the same synapse of one or more neuromodulators with one or more acknowledged transmitters could provide additional pre- and postsynaptic mechanisms for the regulation of neuronal signaling (34). However, the effects of these interactions and their underlying cellular mechanisms are virtually unknown.

Colocalization is not limited to the nervous system. Evidence indicates that TRH and cholecystokinin coexist in a portion of gastrin-secreting cells located in the central mucosa of guinea-pig stomach (27). The importance of the relationship between these two peptides with regard to feeding behaviors and digestive processes has not been determined.

In the decade that followed the announcement of its chemical identity, it was uncertain whether TRH was formed by a nonribosomal enzymatic mechanism or by post-translational processing of a larger precursor protein in ribosomes (28). By 1981, opinion strongly favored the post-translational processing hypothesis because convincing evidence could be adduced only for this alternative. For example, it was demonstrated that a high-molecular-weight protein from frog brain, after chemical and enzymatic treatment, could yield TRH.

A post-translational enzymatic mode of TRH synthesis in amphibians was confirmed using molecular techniques (28). It was determined that frog DNA contained a segment of 478 nucleotides that coded for the amino-terminal region of pro-TRH, a 123-amino-acid precursor containing three copies of the progenitor sequence of TRH (Gln-His-Pro-Gly) flanked by paired dibasic residues and a fourth incomplete copy lacking the C-terminal glycine. A mammalian pro-TRH molecule was later identified in rat hypothalamus as a 255-amino-acid protein containing five copies of the amino acid sequence of the TRH progenitor (28). Human pro-TRH contains six copies (91).

The biosynthesis of TRH is essentially a five-step process (see Fig. 2), beginning with transcription of DNA of the TRH gene to TRH mRNA within the cell nucleus. Transcription is followed by translation of the TRH mRNA to the pro-TRH peptide on the ribosome. The post-translational processing of TRH begins with excision of the progenitor peptides by the action of carboxy peptidases. This is followed by amidation of proline by peptidyl glycine alpha-amidating monoxygenase, the amide moiety being donated by the C-terminal glycine (28). Finally, cyclization of the N-terminal glutamine by glutaminyl cyclase is accomplished (28). Post-translational processing of TRH appears to be restricted to the neuronal perikarya because of lack of TRH progenitor immunoreactivity in axons or terminals of the median eminence or spinal cord (29).

The post-translational processing of pro-TRH also gives rise to a number of other peptides that may have behavioral or physiological activity (6).

The steps of biosynthesis of TRH are similar, if not the same, throughout the CNS. Regulation of synthesis, however, may be more discrete. TRH levels in secretory cells of the hypothalamus must be maintained at concentrations sufficient to ensure functional integrity of the hypothalamic–pituitary–thyroid (HPT) axis, whereas extrahypothalamic levels of the peptide are likely to be responsive to other behavioral and physiological demands.

In the rat, thyroid hormones can regulate hypothalamic TRH production at the transcriptional level by negative feedback. An inverse relationship exists between thyroid hormone status and TRH mRNA levels in medial parvocellular tuberoinfundibular TRH neurons (44); however, this relationship is not observed in other hypothalamic or in extrahypothalamic sites. Thyroid hormones may also regulate translation of the TRH message (44). Whether the effects of altered thyroid states on the transcriptional or translational aspects of TRH synthesis are mediated by intracellular factors such as thyroid hormone receptors or by extracellular neural input is still unclear. It is of interest, however, that a small sequence of DNA flanking the TRH gene is homologous to a sequence of the DNA to which the L-triiodothyronine (T3) receptor binds (45). This sequence is identical to a sequence of DNA near the gene for the beta subunit of TSH, the expression of which is regulated by T3, but is unlike the sequence mediating T3 regulation of growth hormone transcription. On the other hand, it has been shown that the 5¢-deiodinase, which converts thyroxine (T4) to T3, its more active congener, is absent in the periventricular nucleus. This, together with evidence of synaptic connections between axons of epinephrine-containing neurons and perikarya of TRH neurons in the periventricular nucleus, suggests that thyroid hormones may regulate TRH biosynthesis indirectly via central catecholamine pathways (44). For example, cold exposure, which raises central catecholamine levels, elevates cellular TRH mRNA in the paraventricular nucleus even in the presence of elevated thyroid hormone levels (93). This effect on TRH mRNA synthesis is specifically blocked by ethanol; however, ethanol apparently does not block the release of TRH (94).

Regulation of post-translational processing of TRH has not been studied extensively even though it may be rate-limiting for TRH biosynthesis (13). It may also vary by cell type.

The regulation of TRH biosynthesis in extrahypothalamic areas of the CNS is poorly understood. However, it is well established that levels of TRH and TRH mRNA in these areas are not controlled by thyroid hormones (44). TRH biosynthesis varies in different brain areas (6, 44), probably in accordance with specific neuronal stimuli. For example, TRH and TRH mRNA levels are markedly elevated in the hippocampus, amygdala, and pyriform cortex (but not in hypothalamus) after seizures induced by kindling or electroconvulsive shock (42, 72), whereas acute administration of cocaine reduces TRH mRNA in the amygdala and hippocampus (77).

TRH is released into the hypophyseal portal circulation from axonal terminals whose cell bodies are found in the parvocellular division of the paraventricular nucleus (44). Synaptic release of TRH occurs in many parts of the brain. In addition, TRH is co-released with serotonin and substance P from projections of medullary raphe neurons that impinge primarily on cells of the ventral horn and sympathetic lateral column of the spinal cord and perhaps the dorsal horn (27). Similar events may occur in other areas of the CNS (27).

The release of TRH for maintenance of pituitary– thyroid axis homeostasis is regulated by T3, as is TRH synthesis (see above). However, there is provision for override of this negative feedback system (93), probably by central noradrenergic input (30). Furthermore, serotonin (60) and probably GABA (84), somatostatin (49), and corticotropin-releasing factor (49) inhibit hypothalamic TRH release. Dopamine, acting through D2 receptors, appears to produce a general stimulatory effect on the median eminence, eliciting release of both TRH and somatostatin (47). However, dopamine can also directly inhibit TSH secretion (see below and GABA and Glycine, Colocalization in Dopamine Neurons, and General Overview of Neuropeptides).

Synaptic release of TRH in brain has not been studied extensively. It is a complex process, influenced by a variety of neurochemical signals that mediate behavioral and physiological responses to internal and external stimuli. It has been shown that histamine will release TRH from hypothalamic slices whereas somatostatin will do so in brainstem synaptosomes (22). More recently it was demonstrated that the muscarinic agonist pilocarpine will release TRH from slices of the preoptic area of the hypothalamus (17).

TRH is released from cells in the stomach wall into gastric juice by histamine (62) and serotonin (36)). Whether these transmitters also release cholecystokinin, which is thought to be colocalized with TRH in some gastrin-secreting cells of stomach (27), has not been resolved.

Receptors for TRH are found on thyrotroph and mammotroph cells of the anterior pituitary and on neurons throughout the CNS (26, 78). The structural and functional properties of both pituitary and central TRH receptors have been characterized in detail and shown to be similar in structure, binding characteristics, and mechanisms of signal transduction (26, 78). TRH receptors belong to a class of G-protein-coupled receptors with seven membrane-spanning domains and an extracellular N-terminal region containing N-glycosylation sites (81) (see also Signal Transduction Pathways for Catecholamine Receptors and Serotonin Receptors: Signal Transduction Pathways).

Anterior pituitary TRH receptors have been studied extensively in cultured tumor cells originating from thyrotrophs and mammotrophs (26) and more recently in cells transfected with and expressing TRH receptor cDNA (11, 18, 48). All pituitary TRH receptors appear to be structurally identical, though there may be subtle differences in the transduction of hormone secretion between normal anterior pituitary cells and tumor cell models (74).

Pituitary TRH receptors contain saturable, noninteracting, high-affinity binding sites that exhibit strict structural specificity for TRH at all three amino acid positions. The affinity and rate of dissociation of the TRH receptor for its endogenous ligand are highly temperature-dependent: TRH binds with lower affinity and dissociates more rapidly at higher temperatures (26). The rate of dissociation for TRH is also directly related to the length of time of receptor occupancy and may reflect ligand-induced binding site aggregation or receptor internalization (26).

The turnover rate of the pituitary TRH receptor is slow, providing explanation for its equally slow homologous down-regulation (26). The number of pituitary TRH receptors is reversibly decreased by thyroid hormones and increased by estrogen and glucocorticoids. TRH receptors are down-regulated by drugs that elevate cAMP, while receptor affinity is generally reduced by agents that activate protein kinase C (26).

Central TRH receptors exhibit binding characteristics for both 3H-TRH and the high-affinity synthetic TRH analogue 3H-(3-methyL-His2)TRH that are almost identical to those of the pituitary TRH receptor, including the time-dependent biphasic dissociation kinetics indicative of a ligand-induced affinity shift (78). Although heterogeneity of central TRH receptors has been reported, the bulk of evidence, including studies of solubilized highly purified receptors, indicates a single population of high-affinity receptors that are saturable and noninteracting (78).

TRH receptors in all mammals studied thus far, including the human (53), are heterogeneously distributed throughout the CNS. Nevertheless, specific patterns are evident that appear to correlate with species-specific physiological differences (78). Central TRH receptor number seems to be regulated in a conventional way—that is, diminished by TRH and its agonist analogues. It has been shown that dissimilar drugs, including ethanol, delta-9-tetrahydrocannabinol and chlordiazepoxide, also affect TRH receptor binding (79).

Regulation of both the peripheral and central actions of TRH involve intracellular and extracellular enzymatic inactivation of the peptide. In vitro studies have identified several TRH-degrading enzymes, including histidylproline imidopeptidase, prolylendopeptidase, pyroglutamyl aminopeptidase I and II, and thyroliberinase; however, it is generally acknowledged that thyroliberinase and pyroglutamyl aminopeptidase II are primarily responsible for removal of TRH in vivo (65).

TRH released into the hypophyseal portal circulation is degraded rapidly to His-Pro NH2 by thyroliberinase, a highly specific serum peptidase (65). The activity of TRH released into synapses of the CNS is thought to be terminated primarily by pyroglutamyl aminopeptidase II, a high-specificity synaptosomal membrane ectoenzyme (65). Because this enzyme is specific for TRH, a drug (were one available) that would inhibit only this enzyme would elevate the levels of TRH and not the levels of other peptides. Such an indirect TRH agonist might also have diminished endocrine effects. Phorbol ester might appear promising in this regard because it inhibits pyroglutamyl aminopeptidase II, by stimulating protein kinase C-mediated phosphorylation of the enzyme, without blocking its cytosolic counterpart (82). However, the lack of specificity of this reaction would probably limit the use of phorbol ester as an indirect TRH agonist.

The enzymatic breakdown of TRH by cultured cells from the anterior pituitary is regulated by estrogen and thyroid hormones (4). However, the TRH-degrading activity of cultured brain cells is not affected by T3 (4). TRH may inhibit its own enzymatic removal by enhancing translocation of protein kinase C from cytosol to plasma membrane, where it can inactivate pyroglutamyl peptidase II by phosphorylation of the enzyme.

The unequivocal endocrine function of TRH is to stimulate the synthesis and release of TSH from thyrotroph cells of the anterior pituitary gland. Thus, TRH, in concert with thyroid hormones and the inhibitory influences of dopamine and somatostatin from the hypothalamus, controls pituitary TSH synthesis and release (75). T3 decreases hypothalamic TRH and the density of pituitary TRH receptors (see above); it also inhibits TSH subunit gene expression (90) and TSH secretion (75). Dopamine acts directly on anterior pituitary cells through D2 receptors negatively coupled to adenylate cyclase to inhibit TSH synthesis and secretion, while somatostatin likewise reduces TSH release via membrane receptors that modulate adenylate cyclase or voltage-dependent potassium channels (75). TSH initiates the synthesis and triggers the secretion of thyroid hormones, and TRH itself may directly stimulate the thyroid gland to release T4 (2).

TRH stimulates prolactin (PRL) release, but evidence that it is a major physiologic PRL-releasing factor is either species-specific or controversial (46). TRH actions on growth hormone and adrenocorticotropic hormone (ACTH) secretion are complex, being virtually absent in normal subjects and occurring to varying degrees in different pathophysiologic states (75). TRH may exhibit a weak stimulatory effect on follicle-stimulating hormone (FSH) secretion in humans (75).

The basic cellular mechanisms of TRH actions on TSH (87) and PRL (19) secretion have been investigated using cultured pituitary tumor cells. The biphasic secretion of these two hormones has recently been shown to involve G-protein-coupled stimulation of inositol phospholipid turnover. The initial phase of hormone secretion is thought to result from inositol trisphosphate-mediated release of Ca2+ from intracellular stores; the second and more sustained phase results from influx of extracellular Ca2+ via voltage-sensitive channels activated by protein kinase C, which is induced by 1,2-DG (48). Stimulation of TRH receptors on thyrotroph tumor cells also induces the release of arachidonic acid metabolites that may act as additional intracellular messengers (34). In addition to stimulating TSH and PRL release, TRH regulates the transcription of PRL (61) and the post-translational glycosylation of TSH (50) (see also Neuroendocrine Interactions).

The central actions of TRH are myriad, affecting brain chemistry, physiology, and behavior. It has been shown that these centrally induced effects are not mediated by the endocrine effects of TRH but may be harmonious with those endocrine effects (64) and dependent upon the behavioral/physiologic state of the organism (63). As a neuromodulator of several different neurotransmitters, including most prominently dopamine, serotonin, acetylcholine, and the opiates, TRH affects the actions of many drugs that themselves affect these and other neurotransmitters (22, 63).

TRH has been shown to arouse hibernating animals, through a hippocampal mechanism, and to antagonize the sedation, motor impairment, and hypothermia produced by ethanol and other CNS depressants (63). However, unlike other agents that counteract the effects of ethanol such as the GABA antagonists (20), TRH has anticonflict properties, alone and in concert with ethanol and sedative-hypnotic drugs, and may protect against rather than induce seizures (42, 92). The anxiolytic effects of TRH may also be involved in the recently reported attenuating effect of a TRH analogue on alcohol preference in a rodent model of human alcoholism (71). The anticonvulsive actions of TRH and its analogues may prove beneficial in ameliorating alcohol withdrawal symptoms.

TRH counteracts the hypothemia or poikilothermia produced by various drugs and several endogenous neuropeptides, including neurotensin, bombesin, and betaendorphin (22, 63). Its effects alone on body temperature are variable: TRH produces hypothermia in some species, hyperthermia in others, and in some it has no effect. Similarly, TRH produces only weak species-specific effects on nociception but is a potent antagonist of the antinociceptive effect of neurotensin (63).

TRH stimulates locomotor activity by activation of the mesolimbic dopamine system (9, 22, 63). The tripeptide also produces profound stimulation of the cardiovascular and respiratory systems (22, 63) and induces increases in gastrointestinal motility and the volume and acidity of gastric secretion (21) while often suppressing the intake of both food and water (63). These gastrointestinal effects of TRH may play a role in the ulcerogenic actions of TRH, demonstrated in the cold-restraint stress paradigm (21).

The actions of TRH are generally short-lived because of rapid inactivation or biotransformation to less active metabolites (22). In any case, most (if not all) of the CNS effects of TRH can be achieved even when the endocrine targets of the tripeptide have been removed. With peripheral administration, high doses of TRH are required to produce CNS effects because of poor penetration into brain and rapid peripheral inactivation. These limiting properties of the native tripeptide have prompted synthesis of more stable agonist analogues. Analogues with high CNS receptor affinity and greater behavioral, as opposed to endocrine, effects would be desirable; however, the similarity of central and pituitary receptor binding characteristics has made the design of such drugs difficult. Nevertheless, some TRH analogues have been tested as therapeutic agents for disorders of the nervous system (see below).

Initial studies of the effects of TRH as a hypothalamic releasing factor revealed both expected and unanticipated results. Expected findings were as follows: (a) TRH is a potent releaser of TSH in both men and women, with release of TSH occurring at a whole-body threshold dose of 6.25 mg and peak response increasing linearly to a dose of 400 mg TRH (80); (b) the effect of TRH is transient, with peak TSH response occurring about 30 min after administration (80); and (c) the TSH response to TRH is reduced in hyperthyroidism and augmented in hypothyroidism, reflecting the net effect of the potent inhibitory actions of thyroid hormones on TSH release and synthesis (see above). Unexpected findings were as follows: (a) TRH is a potent releaser of PRL in both men and women (32) and (b) women have much greater PRL and slightly greater TSH responses than men.

With the establishment of the basic parameters of the pituitary response to TRH, the development of standardized TRH tests was possible. Two strategies emerged to test the TSH response: In one a submaximal dose of 200 mg TRH is used; in the other a supramaximal dose of 500 mg TRH is given. Probably for historical reasons the 500 mg TRH test became the standard for psychiatric studies. In its simplest version the TRH test requires that a baseline sample of TSH be drawn and then TRH is infused intravenously during 1 min. A second TSH sample is drawn 30 min later. This protocol produces results that are closely correlated with results from more frequent and prolonged sampling (76).

While the TRH test has been a mainstay for providing complete testing of the HPT axis for many years, recent advances in the analytical qualities of TSH assays have reduced its use. The increased sensitivity of TSH assays has enabled detection of hyperthyroid as well as hypothyroid states. Comparison of basal TSH to TRH-induced TSH values has revealed a high correlation between the two measures and raised the question of what role the TRH test should play in HPT axis evaluation (40).

Soon after the advent of the TRH test in endocrinology came the discovery of a blunted TSH response to TRH in some patients with mental depression. This observation, reported in 1972 by Prange et al. (70) and Kastin et al. (37), that the TRH-induced TSH response is reduced in some depressed men and women compared to control subjects, has been replicated by many groups and has become an accepted biological marker in depression (51). Over the past 20 years the significance of this marker has been investigated on several fronts: (a) diagnostic and phenomenological associations, (b) relationship to treatment and prognosis, (c) status as a risk marker, and (d) pathophysiological basis. (For a review of these issues see ref. 51.)

As data accrued it became clear that a blunted TSH response occurs in 25–30% of patients with DSM-III- or RDC-defined major depression. What remains unclear is whether this 25–30% of depressed patients represents a distinct subgroup or the lower end of a single population. TRH testing of many psychiatric patients led to several findings. First, as additional diagnostic groups were studied it became apparent that patients with certain other psychiatric disorders also had significantly higher prevalences of a blunted TSH response than did control populations (57). For example, blunted TSH responses have consistently been reported in patients with alcoholism (15). Less consistent have been reports of blunted TSH responses in patients with borderline personality disorder or panic disorder. Given these findings, it became apparent that the specificity of the blunted TSH response to TRH for depression was poor. This, combined with its low sensitivity, prevented the use of the TRH test as a screening tool for depression.

Separate from standard diagnostic categories, psychopathological features reported to be associated with a blunted TSH response have included suicidality, agitation, and panic (10). However, these findings have not been sufficiently replicated to establish them as reliable correlates of the blunted TSH response.

Another potential area of clinical application for the TRH test is in planning treatment and assessing prognosis. No consistent findings have emerged regarding the use of the TRH test to select treatment. However, there is some evidence that, in depressed patients, those whose TSH responses increase during the process of clinical recovery are less likely to relapse (38).

An important question is whether the blunted TSH response to TRH precedes psychiatric illness. Is it a risk marker for illness or is it a consequence? No studies have adequately addressed this question in depression, but several studies have been completed in individuals who, because of family history, are at high risk for alcoholism. These studies have adduced preliminary evidence of increased prevalence of a blunted TSH response in the nonalcoholic sons of alcoholic fathers (15). If confirmed, this finding would open a new area of investigation for the TRH test as a risk marker for certain psychiatric disorders.

Since the initial reports of a blunted TSH response in depression, the nature of the pathophysiological basis for the fault has been sought. The simplest explanation would be that the blunted TSH response results from increased feedback inhibition by thyroid hormones. Overt hyperthyroidism is uncommon in depression, though transient increases in thyroid hormones are frequent (52, 83) and could account for some instances of blunting (8). The mechanism of the transient hyperthyroxinemia syndrome in depression is unknown; it has been hypothesized to derive from CNS activation either directly to the thyroid gland via sympathetic monoaminergic neurons (56) or indirectly via increased TRH release. In depression, hyperthyroxinemia is usually transient; the blunted TSH response may persist.

Another possible explanation for the blunted TSH response in depression is the hypercortisolemia commonly observed. Cortisol is known to inhibit TSH release, and if this were the cause of a blunted TSH response the phenomenon would unite a pathophysiological change in the HPT axis with one in the hypothalamic–pituitary–adrenal axis. However, a dissociation between these two endocrine faults generally has been found, militating against hypercortisolism as more than an occasional cause of the blunted TSH response (39).

A leading hypothesis to explain a blunted TSH response in depression is hypersecretion of hypothalamic TRH, which should lead to desensitization of pituitary TRH receptors. This hypothesis is attractive because activation of TRH neurons in the periventricular nucleus could be produced by monoaminergic neurons (45, 93) that are thought to be dysfunctional in depression. In support of the hypersecretion hypothesis are reports that TRH is increased in cerebrospinal fluid (CSF) of patients with major depression compared to neurological controls or medically healthy controls (3). However, in neither of these studies was a correlation found between TRH in CSF and the TRH-induced TSH response. The increases in CSF TRH in depression may be related to attempts to compensate for the illness (67).

One study has examined TRH in CSF of abstinent alcoholics and healthy controls and reported no differences between the two groups (73). This suggests that changes in CSF TRH are not simply a function of nonspecific psychopathology.

At present the two leading hypotheses to explain a blunted TSH response in depression are transient hyperthyroxinemia and hypersecretion of hypothalamic TRH. However, it is probable that other factors at the level of the brain or pituitary are also involved in causing the abnormal response (31).

While a blunted TSH response to TRH occurs in some depressed patients, an exaggerated response occurs in others. This latter phenomenon supports the concept of an association of depression with subclinical hypothyroidism. Patients with subclinical hypothyroidism show neither classical signs and symptoms nor reductions in levels of thyroid hormones. The diagnosis depends upon either an elevated basal TSH or an exaggerated TSH response to TRH. It has been reported that 5–15% of depressed patients meet criteria for subclinical hypothyroidism (23), but whether these rates are significantly different from an age- and gender-matched population is uncertain (23). Preliminary data suggest that depressed patients with subclinical hypothyroidism have more cognitive impairment, resistance to antidepressants, and, if bipolar, a greater likelihood of rapid cycling (23).

Finally, it should be mentioned that much of the work on the TRH test in depression occurred prior to the development of highly specific and sensitive TSH assays. Two points need to be made with this in mind. First, it will be important to examine the relationship between basal TSH and TRH-induced TSH release in psychiatric patients to determine if basal TSH can provide sufficient information to obviate the need for the TRH test. Second, definitions of a blunted TSH response that were derived from less specific and sensitive radioimmunoassays are not valid when new TSH assay methods are employed (16).

As described above, shortly after its discovery as a hypothalamic hypophysiotropic hormone, TRH was found to be distributed in many extrahypothalamic sites in brain, in spinal cord, and in other organ systems. The neuroanatomical location and neurochemical actions of TRH suggested that it could be utilized as a therapeutic agent in neurological and psychiatric disorders.

TRH receptors occur in the ventral horn area of the spinal cord. For this reason and because TRH excites anterior motor horn cells, therapeutic trials were initiated in amyotrophic lateral sclerosis (ALS) (12), a disorder characterized by loss of innervation to the anterior horn cells. These studies provided the impetus to administer TRH or its analogues in high doses. Although clear, albeit brief, improvement has been reported in patients with ALS, the application of TRH and its analogues in this disorder is still tentative. TRH has been tried in spinal–cerebellar degeneration, spinal cord injury, and disturbances in consciousness (66). Its efficacy in these conditions is unclear at this time.

The blunted TSH response to TRH in depression occurred while investigating possible therapeutic effects. The initial studies of Prange et al. (70) and Kastin et al. (37) reported that, unlike saline, 500 mg of TRH administered intravenously to unipolar depressed women produced significant improvement in depression ratings that were apparent within 24 hr of administration. Since these initial reports about 30 studies have evaluated the antidepressant efficacy of TRH. Two generalizations seem justified: The number of positive and negative reports is about equal; even in the positive studies, the effects of TRH tend to be short-lived. The reasons for these inconsistent therapeutic effects of TRH are unclear. Preclinical behavioral studies have demonstrated that the peptide, like most antidepressants, potentiates L-DOPA-induced motor activity and decreases immobility in the Porsolt swim test. However, unlike most of its effects on central neurotransmission, which are compatible with an antidepressant-like action, TRH potently enhances cholinergic activity (22, 63), which is, theoretically at least, depressogenic (33). The inconsistency in therapeutic effects, the transient nature of the effects, the lack of availability of TRH analogues for clinical use, and the availability of alternative treatments have delayed further study of TRH as an antidepressant.

In several of the early studies of TRH in depressed patients, improvement was noted in anxiety as well as in depressed mood. The possibility that TRH might have anxiolytic effects also derived support from animal studies indicating that TRH enhances the anticonflict effects of benzodiazepines, barbiturates, and ethanol and, in high doses, produces anticonflict actions itself (85, 89). Very little work has been completed on a clinical anxiolytic effect for TRH. Reductions in tension in nonclinical populations have been reported (88), but studies of patients with a primary anxiety disorder are lacking.

Because of the observation that TRH enhances acetylcholine transmission, high doses of the tripeptide have been tested for possible therapeutic action in Alzheimer's disease. As in depression and ALS, several statistically significant effects were found but with little sustained benefit (57).

TRH is a potent antagonist of the sedation caused by drugs such as barbiturates and ethanol (63) even while potentiating their anticonflict effects. Unlike other sedative–hypnotic antagonists, TRH does not cause seizures and is not anxiogenic; indeed, as noted, it has anxiolytic properties. Because of this unique pharmacological profile, TRH has been studied as an ethanol antagonist in humans. Three studies have been published: two have found evidence that TRH counteracts some of the behavioral actions of ethanol and one that it slightly enhances intoxication (14).

Recently it was reported that an analogue of TRH, TA-0910, produces a reduction in ethanol intake in alcohol-preferring rats (71). The possibility exists that TRH or an analogue would reduce alcohol consumption or alcohol craving, but no relevant clinical data have been reported.

Clearly TRH plays a regulatory role in an elaborately regulated system, the HPT axis. Just as clearly it can directly affect behavioral and physiological events outside that system. Phylogenetically, its functions outside the axis may have developed earlier (30), though they are less thoroughly understood. Given such considerations, how can one conceptualize the diverse functions of TRH?

Elsewhere (64), one of us, with Nemeroff, posited a theory of harmony between the behavioral effects and the endocrine effects of the hypothalamic hypophysiotropic hormones. We illustrated this with several examples, none of which is more striking than that pertaining to luteinizing-hormone-releasing hormone (LHRH). LHRH releases luteinizing hormone and FSH, which stimulate the gonads. LHRH also facilitates mating behavior. TRH lends itself to this general formulation. Through TSH, it causes the synthesis and release of thyroid hormones, which, in turn, not only increase metabolic rate but potentiate the action of ergotropic neurotransmitter activity. TRH, as noted above, exerts a variety of alerting effects. Indeed, it was this array of effects that prompted Kruse (41) and then Metcalf and Dettmar (59) to describe TRH as an ergotropic substance. This notion resonated with the earlier ergotropic–thyrotropic formulation of Hess (25). In this scheme, the organism is always oriented both internally and to its environment somewhere on a spectrum between two extremes. In a complete ergotropic state, the organism is totally engaged in the environment—alert, physically active, and maximally using energy. In a complete trophotropic state, by contrast, the organism is at rest, relatively inattentive to the environment, and restoring itself through the many processes that together are anabolic. In developing this concept, Hess emphasized two points that have become increasingly important. First, recuperation is not merely the absence of excitation; it is an active process. Second, the CNS, prominently the hypothalamus, orchestrates whatever balance of ergotropism–trophotropism may obtain at any time.

Following the concept described above and noting that TRH (and certain other peptides) abbreviated pentobarbital-induced sleep in mice while the tridecapeptide neurotensin (NT) extended it, our group formulated NT as a trophotropic hormone, in balance with the ergotropic tendencies of TRH (63). This concept was later developed in more detail and in the perspective of recent findings (69).

It would appear that the functions of TRH have been conserved during the evolution of simple species to complex ones. Its behavioral role may be older phylogenetically than its role as a regulator of thyroid function. Behaviorally it seems to be an ergotrophic substance (there are probably others) in dynamic balance with trophotropic substances such as NT. The ergotropic behavioral activity of TRH appears to be harmonious with the endocrine actions for which it was named.

This work was supported by grants from the National Institute of Alcohol Abuse and Alcoholism, AA07809 (GAM) and AA08448 (JCG), and from the National Institute of Mental Health, MH33127 (AJP) and MH22536 (AJP).

published 2000